Identification and interpretation of patterns in rocket engine data

Abstract

The goal of our research is to analyze ground test data, to identify patterns associated with the anomalous engine behavior. On the basis of this analysis, it is the task of our project to develop a Pattern Identification and Detection System which detects anomalous engine behavior in the early stages of fault development significantly earlier than the indication provided by either redline detection mechanism or human expert analysis. Early detection of these anomalies is challenging because of the large amount of noise presence in the data. In the presence of this noise, early indication of anomalies becomes even more difficult to distinguish from fluctuations in normal steady state operation.